Enhanced Knowledge Discovery for Social History

Social historians struggle with the problem of describing accurately the attitudes and experiences of underrepresented populations-risking monologizing and so re-objectifying and potentially skewing representations of them, however sympathetic or disinterested new histories attempt to be.

Through augmenting conventional prosopography-”a means of profiling any group of recorded persons linked by any common factor”-with text mining tools and services, we seek to help historical researchers to develop more nuanced perspectives of the vast data collections now available to them. With supporting data synthesis technologies, researchers will be able to sift through accumulated records for patterns with greater refinement and ease, changing parameters to suit research queries as they develop. They will also be able to target key individuals, events, or trends for further in-depth analysis, supporting “thick” or “micro” historical description.

In collaboration with the University of Minnesota, we propose to create a historical search system based on text mining that would supply a powerful research tool for the exploration and discovery of patterns and facts from a variety of historical document collections. Computer enhanced entity extraction is the next logical step for increasing the research value of historical archives. Many historical methods of inquiry begin with quantifiable data that is dispersed within and across texts, and so are difficult for researchers to assemble, except through painstaking and costly labor. Such automated data analysis can also produce faster, more comprehensive, and potentially more nuanced results than researchers can through pen and paper efforts.